EXPERIMENTAL TECHNIQUES AND RESULTS OF INVESTIGATIONS OF AN UNSTEADY FLOWFIELD IN WING WAKES / BOW SHOCK WAVE INTERACTIONS

Size: px

Start display at page:

Download "EXPERIMENTAL TECHNIQUES AND RESULTS OF INVESTIGATIONS OF AN UNSTEADY FLOWFIELD IN WING WAKES / BOW SHOCK WAVE INTERACTIONS"

Laurel Scarlett Pierce
5 years ago
Views:

1 EXPERIMENTAL TECHNIQUES AND RESULTS OF INVESTIGATIONS OF AN UNSTEADY FLOWFIELD IN WING WAKES / BOW SHOCK WAVE INTERACTIONS. Introduction A.M. Shevchenko, A.S. Shmakov Khristianovich Institute of Theoretical and Applied Mechanics SB RAS, 639, Novosibirsk, Russia Shock/Vortex Interaction is one of the fundamental problems of aerogasdynamics, which has been adequately studied neither theoretically nor experimentally. Interference of the vortex with shock waves in the inlet or other elements of the flying vehicle located downstream often leads to vortex breakdown, which, in turn, can deteriorate the lifting capacity of aerodynamic surfaces, lead to inadequate regimes of engine operation and to a drastic increase in heat transfer. Despite the adverse features of this phenomenon, it can be used as one method for improving mixing in the combustor. This phenomenon was first observed in [], where interaction of a vortex core with a system of shock wave at the inlet of a supersonic flying vehicle was examined. In subsequent years, researchers of Russia, Germany, France, and USA studied this problem [2-9]. Experimental data for the Mach number range of.6 to 4 were obtained. Numerical and experimental results on interaction of a supersonic vortex with oblique and normal shocks were reviewed in [6, 8]. It was found in those activities that a conical recirculation zone is formed o to vortex/shock wave interaction. Similarity of boundary-layer separation and vortex/shock wave interaction was established. There is no single opinion, however, concerning the influence of flow swirl on conditions for streamwise vortex breakdown. Thus, it was found on the basis of experimental data [2] that the flow swirl parameter (ratio of the circumferential and axial velocities in the vortex core) is one important characteristic affecting the interaction process. At the same time, the similarity of the interaction process and boundary-layer separation and the results of [, 8, ] show that the interaction process in many cases depends only weakly on the vortex strength and, hence, on the swirl parameter. One specific feature of interaction of a vortex wake with a shock wave is the unsteadiness of this process. It is manifested in experimentally observed fluctuations of gas-dynamic parameters of the flow and the changes in the flow structure and interaction region size with time. Recent numerical results [] showed that the unsteady process may possibly be periodic. A hypersonic interaction of a -tip vortex with a Pitot type inlet has been examined at Mach number of 6 [, 2]. It was obtained that the -tip vortex is the initiator of self-oscillatory process with salient fundamental. Borovoy et al [8] revealed a strong influence of a vortex on heat transfer on a blunted body. Therefore, two sets of experiments were performed to investigate of an unsteady nature of a wake / bow shock wave interaction in a wide Mach number range of 3-6. Present paper described experimental techniques and results of these studies. A.M. Shevchenko, A.S. Shmakov, 28

2 33 Section IV 2. Experimental setup and techniques The experimental study of the interaction of a vortex wake with a bow shock wave were carried out in the supersonic blow-down wind tunnel T-33 of the Khristianovich Institute of Theoretical and Applied Mechanics of the Russian Academy of Sciences. It is an intermittent blowdown facility with the test section dimensions of m. The wind tunnel is capable of producing a Mach number range from.75 to 7 and unit Reynolds numbers in the range of ( 8) 6 per meter. M M 8 α 6 35 bow shock generator Front view y O r x R p p Fig.. Schematic of the experimental setup (not to scale, all dimensions are in mm). Figure illustrates the schematic representation of the experimental setup. A -tip vortex is generated by an unswept semispan slender. A hexahedral airfoil has a chord length of 8 mm, a half-angle of 8 deg. The was mounted on the test section floor at the tunnel centerline. The vortex generator could be set at different angles of attack in the range from -3 to 3 deg. The bow shock wave is generated by a cylinder obstacle with a radius R = mm. The cylinder is placed 35 mm downstream of the trailing edge of the. Shadowgraphs of the flow were taken using a novel visualization technique [3]. Instead of Foucault knife we used a plate made of phototropic glass, further - AVT (adaptive visualizing transparency). Under the influence of focused radiation the AVT material darkens pro rata to intensity. The sensing radiation which is being deflected even at small disturbances on very small angles goes through non-black-out area of AVT and reveals in the shadow picture as more bright part. The great advantage of this technique is an opportunity to visualize weaker perturbations against a background of stronger ones. The scheme as being self-adjusting one essentially simplifies the process of the experiment. The shadowgraph images were carried out using a spark light source with 2.5 microsecond exposure times. The images were recorded at a rate of 8 frames per second. Three piezoelectric fast-response pressure transducers were used to measure of pressure pulsation on the face of the cylinder. The transducers were placed on distance.75r from the center of the cylinder (Fig. ). The amplified output from the transducer was digitized using an AD converter at a rate of khz. Time-averaged pressure distribution on the cylinder butt was measured as well. The pressure ports were located along a line marked as r in Fig.. At M = 6 experimental data were obtained for angles of attack up to 2 deg with the step of 5 deg. At M = 3 and 4 the experiments were performed for angles of attack up to deg with the step of 2.5 deg. 2

Three 5-hole conical pressure probes were used to measure local Mach number, Pitot pressure and flow angularity in the core of the -tip vortex and its vicinity.

3 3. Flowfield in a wake behind of the Fig. 2. Laser sheet images for x/b=.5 (left) and x/b=3. (right): (a) α=5 deg; (b) α= deg; (c) α=5 deg; (d) α=2 deg. In preliminary experiments a flowfield behind of the was examined with a laser sheet imaging technique and multi-hole pressure probe technique []. Three 5-hole conical pressure probes were used to measure local Mach number, Pitot pressure and flow angularity in the core of the -tip vortex and its vicinity. Each probe has an outer diameter of 3 mm. Quantitative flowfield measurements and laser sheet visualization were performed at two cross-sections at the distance of 2 mm and 24 mm downstream from the trailing edge. The experiments were performed at Mach numbers of 3 and 4, and corresponding unit Reynolds numbers of 36 and 56 million per meter. Vortex generator was mounted at angles of attack α = 5,, 5 and 2 deg. Figure 2 shows typical images obtained with the laser sheet technique. In these figures, the vortex core is seen as a black region. In the near wake (x/b =.5) the laser sheet images has revealed a complicated flowfield structure with formations of the bow shock, inner shocks and expansion waves. The observed flowfield is similar to the flow topology over a rectangular [9]. In the far wake (x/b = 3) only the bow shock and the vortex sheet which roll-up of the vortex core are visible. Inner shocks and expansion waves are dissipated. In these figures, the vortex core is seen as a black region. Figure 3 show typical results of 5-hole pressure probe measurements. These measurements yielded detailed data on the fields of total pressure, Mach number, and Mach number components onto the axes of the flow-fitted coordinate system. As is well known in the vortex core the radial component of velocity is small as compared with the circumferential velocity. Thus it is easy to show that crossflow Mach number defined as is a good estimate of circumferential Mach number and is estimation of swirl ratio. A comparison of visualization and probe measurements results showed that the vortex core, which looks as a dark region in the laser sheet images, is characterized by a decrease in total pressure, Mach number, and axial component of velocity; thus, a wake-type profile is formed. The maximum values of the circumferential velocity as well as swirl ratio correspond to the edge of the vortex-core. 3

4 Section IV.. From Laser Sheet image: h_core y / b My Mz Y / b τ = Mzy / Mx z / b = z / b =.5 z / b =. z / b = z / b τ (a) (b).. y / b z/b= 5 z/b=.375 z/b=-5 z/b=.75 z/b=.325 z/b=-.75 z/b=.75 z/b=.425 z/b=-.75 y / b z/b= 5 z/b=.375 z/b=-5 z/b=.75 z/b=.325 z/b=-.75 z/b=.75 z/b=.425 z/b= Mz My (c) (d) Fig. 3. Five-hole pressure probe data through the vortex core for M=4, α= deg: (a) crossflow vectors; (b) swirl ratio; (c) horizontal Mach number; (d) vertical Mach number. These experiments revealed many vortex core characteristics commonly found in previous studies of supersonic -tip vortices [4] and leading edge vortices [5], including an asymmetric swirl distribution and significant total pressure and total Mach number deficits. The spatial scale, Mach number deficit and total pressure loss of the vortices were observed to increase with angle of attack. 4

4. The interaction at Mach numbers of 3 and 4 In experiments without the a well known flowfield with a detached bow shock was observed.

Time-averaged pressure distribution is typical with a maximum in the cylinder centre (see Fig. 7).

The vortex wake behind the radically changes a flowfield in front of the cylinder. Figures 4-6 shows typical shadowgraphs of the flowfield generated during the interaction.

In this case the interaction area is bounded by complex pulsing system of the shock waves interacting with each other.

5 4. The interaction at Mach numbers of 3 and 4 In experiments without the a well known flowfield with a detached bow shock was observed. The flowfield is steady without of any noticeable variation of the distance between the bow shock and the cylinder face. Time-averaged pressure distribution is typical with a maximum in the cylinder centre (see Fig. 7). The standard deviation of pressure fluctuations did not exceed of % of mean free-stream Pitot pressure. The vortex wake behind the radically changes a flowfield in front of the cylinder. Figures 4-6 shows typical shadowgraphs of the flowfield generated during the interaction. Follo modes of interaction have been revealed. At α = and 2.5 deg the pulsing regime was observed. In this case the interaction area is bounded by complex pulsing system of the shock waves interacting with each other. The shape of the bow shock wave and the size of the interaction region change during one test (Fig. 4). Figure 7 shows time-averaged pressure distribution on the cylinder butt. It displays that the pressure decreases in the central part of the cylinder while the angle of attack increases up to 5 deg. (a) M=4, α= (b) M=4, α=2.5 deg Fig. 4. Shadowgraphs of the flow during bow shock / vortex interaction. Pulsing bow shock. (a) M=3, α=5 deg (b) M=4, α=5 deg (c) M=4, α= deg Fig. 5. An interaction mode of "the closed cavity" type. Fig. 6. An interaction mode of "the opened cavity" type (M=4, α= deg). 5

6 Section IV The increase in an angle of attack leads to change of a mode of interaction. Thus in Fig. 5 the displayed flowfield is similar to flow in a closed cavity. This type of the interaction region is wellknown on previous study of the normal shock/vortex interaction [7, 9, ]. In this case the interaction zone is bounded by a round-nosed conical shock wave. The interaction region consists of two zones: a central zone containing the distorted vortex and an outer supersonic zone bounded by the conical bow shock wave. The weak waves in this (outer) zone may be visible in Fig 5. At the further increase in an angle of attack the interaction area extends upstream from the cylinder. And finally, it reaches the vortex generator. Typical shadowgraph for this case is shown in Fig. 6. It is seen that the flowfield is similar to flow in an open cavity. It should be noted that, at α = deg an examination of multiple spark shadowgraphs has revealed an unsteady character of the interaction. But in contrast to the case α = 2.5 deg the shape of the conical shock does not change. P / P t.5.5 Without α= o α=2.5 o α=5 o α=7.5 o P / P t.5.5 without α= o α=2.5 o α=5 o α=7.5 o α= o.5.5 r / R (a) M=3 (b) M=4 Fig. 7. Time-averaged pressure distribution on the cylinder butt.5.5 r / R Figure 7 shows pressure distribution on the cylinder butt. As is seen the vortex wake induces pressure loss at cylinder face. In the central part of the cylinder the pressure not exceed 25% of the Pitot pressure in free-stream. It corresponds to the results received in previous experiments and is caused, apparently Pitot pressure loss in the vortex wake behind the [, 4]. The measured pressure fluctuations at the butt of the shock-wave generator confirm the extremely unsteady character of interaction. Figure 8 shows normalized power spectra density distribution at the cylinder. At the same time, the spectral analysis has not revealed significant peaks in the normalized power spectral density distribution. PSD, db/hz Without α= o α=2.5 o α=5 o α=7.5 o F, Hz PSD, db/hz F, Hz (a) M=3 (b) M=4 Fig. 8. Power spectra density distribution on the cylinder butt free stream α= α=2.5 α=5 α=7.5 α= 6

7 P / Pt x/d=3.5, M=3 x/d=3.5, M=4 x/d=4.8, M=3 x/d=4.8, M=4 σ P / P t x/d=3.5, M=3 x/d=3.5, M=3 x/d=3.5, M=4 x/d=3.5, M=4 cylinder in free stream x/d=4.8, M=3 x/d=4.8, M= α o Fig. 9. Time-averaged pressure at the center of cylinder vs angle of attack α o Fig.. Standard deviation of pressure vs angle of attack Figures 9 and plots pressure at the center of the cylinder and standard deviation of pressure pulsations vs the angle of attack respectively. It is seen that at pulsing mode (α = and 2.5 deg) the pressure and the standard deviation decrease while the angle of attack decreases. At α > 5 deg these both quantities does not depends on the angle of attack of the. Thus, at a pulsing mode the pressure decreases with growth while the angle of attack increases. When the shape of the bounded shock does not change or the interaction region extends upstream to the the pressure distribution does not depends on the angle of attack. The measured pressure fluctuations at the butt of the shock-wave generator confirm the extremely unsteady character of interaction. At the same time, the spectral analysis has not revealed significant peaks in the normalized power spectral density distribution. 5. The interaction at Mach number of 6 Figure illustrates typical shadowgraphs of the flowfield at the Mach number of 6. This figure clearly indicates the oscillating process during the hypersonic interaction of a wake with a bow shock wave. Only at a zero angle of attack the interaction region of a finite size was observed. Otherwise (α = 5 2 deg) it pulsed upstream from the cylinder up to the. The pressure measurements confirm the occurrence of a pulsing interaction process. The normalized power spectral density distribution displays sufficiently high power of the pressure fluctuations. Figure 2a indicates a fundamental with a sufficiently high power and a frequency of about f =38-45 Hz associated with the occurrence of a self-oscillatory process during the interaction at α= 2 deg. Subharmonics also are visible in this figure. At α = and 5 deg the pressure fluctuation power is sufficiently high. But the normalized power spectral density distribution has no peaks. 7

Section IV (a) α=5 deg (b) α= deg Fig.. Shadowgraphs of the flow during bow shock / vortex interaction P / P t.5.5 Without α= o α=5 o α= o α=5 o α=2 o.5.5 (a) Power spectral density distribution (b) Time-averaged pressure distribution Fig.

It is seen that the pressure at central part of the cylinder does not depend of an angle of attack. It corresponds to the case while the interaction region extends upstream to up to the.

The experiments revealed a highly unsteady processes during the interactions. Global reorganization of flowfield structure with a high level of pressure fluctuations was observed.

8 Section IV (a) α=5 deg (b) α= deg Fig.. Shadowgraphs of the flow during bow shock / vortex interaction P / P t.5.5 Without α= o α=5 o α= o α=5 o α=2 o.5.5 (a) Power spectral density distribution (b) Time-averaged pressure distribution Fig. 2. Pressure on the cylinder butt Figure 2b shows a time-averaged pressure distribution at Mach number of 6. It is seen that the pressure at central part of the cylinder does not depend of an angle of attack. It corresponds to the case while the interaction region extends upstream to up to the. 6. Conclusion The interaction of a vortex wake with a bow shock wave was studied at a Mach numbers range from 3 to 6. The experiments revealed a highly unsteady processes during the interactions. Global reorganization of flowfield structure with a high level of pressure fluctuations was observed. At low angles of attack the experiments has detected an interaction process with a pulsing bow shock wave. The self-oscillatory process was revealed at Mach number of 6. Acknowledgements The work was supported by the Russian Foundation for Basic Research, grant No r / R 8

9 REFERENCES. Zatoloka V.V., Ivanyushkin A.K., Nikolayev A.V. Interference of vortices with shocks in airscoops. Dissipation of vortexes // Fluid Mechanics, Soviet Research, 978, Vol. 7, P Delery J., Horowitz E., Leuchter O., Solignac J. Fundamental Studies on Vortex Flows // La Recherche Aerospatiale, (English ed.), 984, P Glotov G.F. Interference of a vortex core with shock waves in a free stream and nonisobaric jets // Uchenie zapiski TsAGI, 989, vol. 2, No. 5, (in Russian). 4. Ivanyushkin A.K., Korotkov Yu.V., Nikolayev A.V. Some features of an interference of shock waves with aerodynamic wake // Uchenie zapiski TsAGI, 989, Vol. 2, No. 5, P (in Russian). 5. Cattafesta L. and Settles G. Experiments on shock/vortex interaction // AIAA Paper, 992, No Delery J.M. Aspects of vortex breakdown // Progress in Aerospace Sciences, 994, Vol. 3, P Kalkhoran I.M., Smart M.K., Betti A. Interaction of a supersonic tip vortex with a normal shock // AIAA Journal, 996, vol. 3, No. 34, P Kalkhoran I.M., Smart M.K. Aspects of shock wave-induced vortex breakdown // Progress in Aerospace Sciences, 2, Vol.3, P Borovoy V.Ja., Kubishina T.V., Skuratov A.S., Yakovleva L.S. Vortex in a supersonic flow and its influence on a flowfield and heat transfer of the blunted body // Mechanica zhidkosti i gaza, 2, No. 5, P , (in Russian).. Shevchenko A.M., Kavun I.N., Pavlov A.A., Zapryagaev V.I. Review of ITAM experiments on shock / vortex interaction // European Conference for Aerospace Sciences, Moscow, July 4-7, 25, Paper No. 2.7., P Zheltovodov, A.A., Pimonov, E.A., Knight, D.D. Numerical modeling of vortex/shock wave interaction and its transformation by localized energy deposition // Shock Waves, 27, Vol. 7, P Shevchenko A.M., Kavun I.N., Pavlov A.A., Zapryagaev V.I. Visualization of -tip vortices and of an unsteady flowfield in shock / vortex interaction // 2th Intern. Symp. on Flow Visualization, Gottingen, Germany, September -4, 26, ISBN , Paper No. 29, P Golubev M.P., Pavlov A.A., Pavlov Al.A. Use of phototropic materials as visualizing transparency in shadow devices // 9th International Scientific and Technical Conference Optical Methods of Flow Investigation, Moscow June 27-29, Smart M.K., Kalkhoran I.M., Bentson J. Measurements of supersonic tip vortices // AIAA Journal, 995, Vol. 33, No., P Brodetsky M.D., Shevchenko A.M. Experimental investigation of the supersonic flow on a lee side of a delta // Thermophysics and Aeromechanics, 998, Vol. 5, No. 2, P

Separation in three-dimensional steady flow. Part 3: TOPOLOGY OF SOME REMARKABLE THREE-DIMENSIONAL FLOWS

Separation in three-dimensional steady flow Part 3: TOPOLOGY OF SOME REMARKABLE THREE-DIMENSIONAL FLOWS H. Werlé. Onera Separation on a blunt body Separation on a blunt body Two-vortex structure. Skin